bm F(t) bf + dm- df as the sex ratio at any time when that ratio is not changing, that is, the ultimate sex ratio (11). This easily obtained result is instructive. The sex ratio in the population will tend to be lower if the male death rate is higher than the female; it will tend to be higher if the sex ratio at birth bmlbf is higher. The latter is not for the moment to be regarded as a policy variable, but it will become one when parents are able to control the probability of a birth being a boy. To find the consequences of such control we would need a model in which the number of births is determined by parents of both sexes. Our present model is called female dominant to signify its restriction; we could easily modify it for male dominance or, what is more realistic, for mixed dominance. To ascertain the consequences of varying sex-ratio at birth on the supply of brides and grooms in the next generation, and hence the ultimate consequences for the birth rate, should be within reach of investigation. If the crude birth and death rates now prevailing among men and women in the United States were to continue they would lead to a low sex ratio. The crude death rate for males in 1966 was 1000dm = 10.87, while the crude rate for females was \QQQd f = 8.11; birth rates were 10006m = 19.05 and 10006.- = 17.61 respectively. We would have for the ultimate sex ratio 19.05 = F 17.61 + 10.87 - 8.11 However, the crude rates will not persist if the age-specific rates do, and we want also to see the outcome in the latter case. This is obtained by entering the corresponding intrinsic rates and provides 2079 = ' F 19.30 + 11.10 - 9.60 a very different result. Apparently with persistence of the age-specific rates the female population becomes older, and this raises its crude death rate and shifts the balance of the sexes towards equality. Births by Age of Father The simpler one-sex model recognizing age can be applied to males. Confining it to females omits half the data; in the face of sampling and some other kinds of error we have twice as much information both on mortality and on ages of parents when we use the male data as well as the female. Thisore than offsets the effect of simple parity. Whelpton (8) did the basic work on parity, and Karmel (9) suggested the importance of nuptiality. Oechsli's (7) recent calculations show the importance of both.